Claude Opus 5 Tops TxBench Preclinical Pharmacology at 68.7%, Gemini 4 Argon Strong in Antibody Evals
kenbwork · x · 2026-10-23
benchmarks.bio published TxBench biomedical agent benchmark results:
- TxBench-Preclinical-Pharmacology (100 tasks built from real experimental data, graded deterministically on whether the agent recovers the original biological result): Claude Opus 5 with Claude Code ranks first at 68.7%, Gemini 3.8 Flash (Pi) second at 66.0%, Grok 4.6 third at 64.7%.
- The poster adds that Gemini 4 Argon ranks 4th (63.0%) on the antibody eval and 3rd (64.7%) on preclinical pharmacology.
- Scores are reported per model-harness pair since harness choice substantially changes results, and are not comparable across benchmarks.
A useful capability reference for models as drug-discovery analysis agents.
Related event: Gemini 4 Argon Ranks High in TxBench Biomedical Agent Benchmarks(2 posts)→
More from Models
- Burkov Shows LLM Copyright Guardrail Applies Only to Content Form, Not Substance — burkov · 2026-10-23
- Google Cloud's Physical-Engineering Last Exam: all frontier models score under 50% — Afinetheorem · 2026-10-23
- LLM copyright guardrail absurdity: refuses to scrape a book but happily zips it for you — burkov · 2026-10-23
- 'Wait' markers are inherited from training data, not invented by RL, HKUST study finds — huawei-noah · 2026-10-23
- Study: Claude Opus 4.8 beats human experts on every measure as peer reviewer of 124 Harvard papers — emollick · 2026-10-23
- AI power user: back in the groove with Claude, still wishes they'd used Perplexity instead — eschadiol · 2026-10-23